Citekey: @blunt02factor

Blunt A and Yang B (2002). “Factor Structure of the Adult Attitudes Toward Adult and Continuing Education Scale and its Capacityt O Predict Participation Behavior: Evidence for Adoption of a Revised Scale.” Adult Education Quarterly, 52(4), pp. 299-314. <URL:>,, <URL:>.


Briefly, this paper trim the widely used AACES from 22 items to 9 items and further checked the revised instrument’s validity and reliability and predictive strength.


This study subjects the AACES to a rigorous examination. Confirmatory factor analysis revealed a three-dimensional structure of attitude with three factors, Enjoyment of Learning, Perceived Importance, and Intrinsic Value. A substantial number of the AACES items were judged extraneous, and a nine-item Revised Attitude Toward Continuing Education Scale (RAACES) was constructed. (p. 2)

Attitudes Toward Continuing Education Scale (AACES) (Darkenwald & Hayes, 1988; Hayes & Darkenwald, 1990) is one instrument purported to be a valid and reliable scale to measure attitudes toward adult education (p. 3)

Although the AACES was introduced more than 10 years ago, a rigorous and independent replication and assessment of the scale’s factor structure and psychometric properties has not been published. (p. 3)

criticized by Schwab (1980) more than two decades ago, the paucity of welldeveloped psychometric instruments frequently leads researchers to use whatever instruments are available. (p. 3)

The development of the AACES drew on Rokeach’s (1968; Rokeach & Kliejunas, 1972) two-attitude theory of attitude toward a psychological object and attitude toward the situation in which the object is encountered (Darkenwald & Hayes, 1988; Hayes & Darkenwald, 1990). (p. 3)

In the first analysis, published in IJLE (Darkenwald & Hayes, 1988), the authors reported a factor analysis that yielded five factors with eigenvalues greater than 1.0 prior to rotation. Because the first unrotated factor with an eigenvalue of 7.2 was dominant, the authors declared the scale to be unidimensional. They further stated, “To double-check the foregoing evidence of unidimensionality, two, three, four and five factor oblique and orthogonal factor analyses were computed. None yielded conceptually meaningful results” (p. 199). In summary, the authors reported that their analyses failed to confirm Rokeach’s (1968; Rokeach & Kliejunas, 1972) theoretical two-attitude model and attitude toward continuing education was unifactorial. Unfortunately, the authors and the editors and reviewers of IJLE failed to recognize that the reported factor analysis results actually confirmed AACES to be multifactorial. In contradiction with recognized interpretations of eigenvalues (see, e.g., Comrey & Lee, 1992), the authors interpreted the presence of a number of eigenvalues greater than 1.0 to be inconsequential when they claimed that the scale was one-dimensional (p. 202). (p. 4)

In their second report, published in AEQ, Hayes and Darkenwald (1990) (p. 4)

The three-factor solution thought to best fit the data was arrived at by orthogonal rotation, and the factors were labeled Enjoyment of Learning Activities (7 items), Importance of Adult Education (9 items), and Intrinsic Value of Adult Education (5 items). (p. 4)

In both published reports on the development of AACES, the authors ignored prior evidence indicating that attitudes toward adult education were most likely multifactorial. The widely published work of Boshier (see Boshier & Collins, 1985) on the EPS, which has its origins in Houle’s (1961) hypothesized tripartite conceptualization of motivational orientations (the same research cited by the developers of AACES to ground their research in existing adult education literature), confirms a six-factor model of motivational orientations. (p. 4)

The study incorporated two objectives. The first objective was to implement a rigoroustestoftheAACES’factorstructureusingtheconfirmatoryfactoryanalytic technique developed by Jöreskog (1969). (p. 5)

The second objective was to examine the predictive validity of the AACES using the more powerful approach of structural equation modeling. (p. 5)

Participants were 458 adult learners enrolled in a variety of adult education programs offered by a western Canadian university and other adult education agencies (p. 5)

RESULTS (p. 7)

Measurement Models for Attitudes and Participation Behavior (p. 7)

Taken together, the results obtained indicated that psychometric refinement of the three subscales was needed with regard to factor content and the number of items. (p. 8)

Therefore, it was concluded that the five-item index formed an adequate measurement model for a single construct of adult education participation behavior. (p. 9)

Having established the superiority of the three-factor model, the analysis focused on a search for an adequate measurement model for the construct of attitude within the AACES item pool. In modifying the AACES, items were deleted one at a time based on modification indices (MI) that pointed to individual items having either a complex loading pattern or being correlated with measurement error. An item with a complex loading pattern, or a passenger item, is an item with a high factor loading on the primary factor as well as a high loading on at least one other secondary factor. Items were discarded from the scale one at time, with the item having the largest MI on at least one secondary factor being selected for deletion. (p. 9)

In other words, more than 90% of attitudinal variations measured on the modified scale could be explained by the three-factor structure of attitudes generated from the nine-item instrument. (p. 10)


To assess the predictive validity of the three-attitudinal constructs of the RAACES, we used a structural equation modeling technique to examine the relationship between attitudes and participation behavior (PBI). (p. 12)

For the three attitude factors, we think that Enjoyment of Learning reflects an affect component, whereas Perceived Importance and Intrinsic Value of Adult Education reflect beliefs and values held about adult education (Hayes & Darkenwald, 1990). (p. 12)

Figure 2 presents this causal model of attitude and participation behavior with estimates of the structural coefficients for the overall sample. All the path coefficients obtained were statistically significant (p < .05), and the fit of the model to the data was acceptable, χ2(69) = 592.39, p < .01; GFI = .86, CFI = .71. Almost 90% of the variance and covariance of the measures on attitude and participation were explained by the model. (p. 13)

The results also indicate that participation in adult education functioned effectively as a construct, with 14% of its variance attributable to the assessed attitudes toward continuing education. Participation behavior was directly influenced by Enjoyment of Learning (.37), which in turn was almost equally affected by Perceived Importance and Intrinsic Value of Adult Education, their path coefficients being .39 and .36, respectively. Although the impact of Intrinsic Value on Enjoyment of Learning was moderate (.36), it had a much stronger influence on Importance of Adult Education (.74). In summary, the results obtained demonstrate that all three factors of adult attitude toward continuing education have significant and interpretable influences on participation behavior. (p. 13)


Revised Adult Attitudes Toward Continuing Education Scale Items

  1. Continuing education is mostly for people with little else to do.
  2. I dislike studying.
  3. Successful people do not need continuing education.
  4. I am fed up with teachers and classes.
  5. Money spent on continuing education for employees is money well spent.
  6. Continuing my education would make me feel better about myself.
  7. I enjoy educational activities that allow me to learn with others.
  8. Continuing education is an important way to help people cope with changes in their lives.
  9. Continuing education helps people make better use of their lives.
  • Enjoyment of learning: 2, 4, 7
  • Importance: 3, 1, 5
  • Intrinsic value: 9, 6, 8
Blog Logo

Bodong Chen



Crisscross Landscapes

Bodong Chen, University of Minnesota

Back to Home